Host‐related and environmental factors influence long‐term ectoparasite infestation dynamics of mouse lemurs in northwestern Madagascar

Parasite infestations depend on multiple host‐related and environmental factors. In the case of ectoparasites, which are exposed to the environment beyond the host, an impact of climate, expressed by seasonal or yearly variations, can be expected. However, long‐term dynamics of ectoparasite infestations are rarely studied in nonhuman primates. We investigated the yearly variations in ectoparasite infestations of two small primates, the gray (Microcebus murinus) and the golden‐brown (Microcebus ravelobensis) mouse lemur. For a more comprehensive evaluation, we also analyzed the potential effects of yearly and monthly climatic variation (temperature, rainfall) in addition to habitat, host sex, age, species, and body mass, on ectoparasite infestation. Individuals of both host species were sampled in two study sites within the Ankarafantsika National Park in northwestern Madagascar during several months (March–November) and across 4 years (2010, 2011, 2015, 2016). Our results show significant monthly and yearly variations in the infestation rates of three native ectoparasite taxa (Haemaphysalis spp. ticks, Schoutedenichia microcebi chigger mites, Lemurpediculus spp. sucking lice) and in ectoparasite species richness in both mouse lemur species. In addition, significant impacts of several host‐related (species, sex, body mass) and environmental factors (habitat, temperature, rainfall) were found, but with differences in relevance for the different parasite taxa and partly deviating in their direction. Although some differences could be attributed to either permanent or temporary presence of the parasites on the host or to ecological differences between the host species, the lack of specific knowledge regarding the life cycle and microhabitat requirements of each parasite taxon precludes a complete understanding of the factors that determine their infestation dynamics. This study demonstrates the presence of yearly and monthly dynamics in lemur–parasite interactions in tropical, seasonal, dry deciduous forests in Madagascar, which call out for broad ecological long‐term studies focusing both on primate hosts and their parasites.


| INTRODUCTION
Variations in the parasite load of nonhuman primates have been studied in various host species in recent decades in the context of assessing primate health, understanding host specificity, parasite diversity, host-parasite coevolution, or the impact of parasites on host fitness and population dynamics (e.g., Altizer et al., 2007;Chapman et al., 2005;Pedersen et al., 2005;Whiteman & Parker, 2005). It is generally accepted that a broad suite of hostrelated factors and the environment beyond the host can impact parasite infestations.
Among the most widely stated host-related determinants of parasite richness and prevalence are species, sex, age, and body condition (e.g., Clough et al., 2010;Durden et al., 2021;Ishii et al., 2017;Klein et al., 2018). The impact of host species on parasite infections can result from differences in ecology and sociality. For example, high host population density can lead to increased interactions between individuals and thus facilitate the transmission of parasites or pathogens (Stringer & Linklater, 2015).
Similarly, the social structure of hosts can play a role, as gregariousness can strongly influence the risk of infection (Patterson & Ruckstuhl, 2013). Conversely, social species executing allogrooming might, in turn, have a lower ectoparasite load (Akinyi et al., 2013). Due to differences in hormone levels influencing immunity or to sexspecific behaviors and ranging patterns, host infection susceptibility might vary between the sexes (Klein, 2004;Rodriguez et al., 2015).
For instance, Zohdy et al. (2017) found that brown mouse lemur (Microcebus rufus) males had significantly more lice than females, which could be due to an immunosuppressive effect of some hormones, like testosterone, and higher encounter rates of male hosts. Rodriguez et al. (2015) found that Microcebus griseorufus males had higher tick infestation prevalences probably due to longer foraging times compared to females. Since older individuals could show immunosenescence (Haberthur et al., 2010;Palacios et al., 2007;Zohdy, 2012), or juveniles could have less developed immune systems (Attanasio et al., 2001;Foerster et al., 1997), age has been argued and shown to impact parasitic infections. In a study on wild Japanese macaques (Macaca fuscata), Ishii et al. (2017) found that juvenile hosts harbored more louse eggs than adult macaques. As a consequence of nutritional stress, hosts in poorer body conditions are often more susceptible to infections than individuals in better conditions (Beldomenico & Begon, 2010;Kiene et al., 2020;Wikel, 1982). However, Arneberg (2002) found parasite species richness to be positively related with host body mass, probably linked to higher transmission rates and the establishment of more parasite species in those hosts that have higher movement rates and food intake.
Suggested environmental impacts on parasite infestations are habitat type, condition and disturbance, temperature, humidity, rainfall, and seasonality (e.g., Blersch et al., 2021;Chapman et al., 2006;Gillespie & Chapman, 2008;Kiene et al., 2021). For example, habitat fragmentation as a predictor for habitat integrity was shown to impact parasite ecology and infection dynamics.
However, previously documented effects of environmental modification were contradictory, since both, higher (Raharivololona & Ganzhorn, 2009;Schwitzer et al., 2010) and lower levels (Kiene et al., 2020;Renwick & Lambin, 2013) of parasite infections have been found in hosts from fragmented and disturbed ecosystems compared to hosts from pristine habitats. These contradicting results, however, may at least be partly explained by the different parasite types they focused on.
Also, climate and its seasonal variations might impact parasites directly or indirectly via their hosts. However, long-term dynamics of parasite infections have only rarely been studied, and the underlying causes driving these dynamics are still under debate. For example, Blersch et al. (2021) found that in wild vervet monkeys (Chlorocebus pygerythrus) dwelling in a semiarid riverine woodland, precipitation negatively affected infections with gastrointestinal nematodes of the genus Protospirura, while the maximum daily temperature was positively related to infections with a nematode from the genus Trichostrongylus. Rodriguez et al. (2015) found that in three noncontiguous dry forests in southwestern Madagascar, ticks parasitized gray-brown mouse lemurs (M. griseorufus) only during the dry season. Moreover, the development and survival of temporary ectoparasites like ticks and chigger mites can depend on humidity, rainfall, and temperature (Berger et al., 2014;Cumming, 2002;Sasa, 1961;Wall & Shearer, 1997). Finally, even though permanent ectoparasites (e.g., sucking lice) should be better protected against environmental fluctuations by their continuous association with the host, they may still be negatively affected by aridity or high ambient temperatures (Kiene et al., 2020;Moyer et al., 2002;Wall & Shearer, 2001a).
On the island of Madagascar, the man-made transformation of natural habitat and climatic fluctuations are severely impacting the entire flora and fauna (Ingram & Dawson, 2005;Vieilledent et al., 2018). The country is inhabited by many endemic vertebrate species, including a large radiation of lemurs (Mittermeier et al., 2010), which host a broad suite of parasite species (Ehlers et al., 2019;Irwin & Raharison, 2009;Kiene et al., 2020Kiene et al., , 2021Klein et al., 2018;Springer & Kappeler, 2016;Zohdy & Durden, 2016). Changes in the distribution of these parasite communities in connection with climate change have already been predicted for the near future by model calculations (Barrett et al., 2013). Since most lemurs are already threatened due to ongoing deforestation and forest fragmentation (www.redlist.org; Schwitzer et al., 2014), it is more important than ever to understand their long-term interconnectivity within their native ecosystems. The investigation of ectoparasites is challenging because it requires direct host examinations, so most multi-year parasitological studies have focused on infections with gastrointestinal parasites (Petrzelková et al., 2010;Rondón et al., 2017), which can be studied noninvasively using fecal samples. Since ectoparasites might be notably and continuously exposed to the environment beyond their hosts, they are particularly interesting study models with respect to their responses to changing environmental conditions.
In this study, we investigate the yearly variation in ectoparasite infestations for two small primate species, the gray (Microcebus murinus) and the golden-brown mouse lemur (Microcebus ravelobensis) from two sites of dry deciduous forest within the Ankarafantsika National Park in northwestern Madagascar. These sites differ largely in forest structure, forest composition, soil type, and proximity to surface water and, therefore, can be considered different habitats (Chanu et al., 2012;Rendigs et al., 2003;Sehen et al., 2010). Both host species have been studied in a long-term project focusing on their socioecology and population dynamics (Chanu et al., 2012;Henkel et al., 2019Henkel et al., /2020Radespiel et al., 1998Radespiel et al., , 2003Radespiel et al., , 2021Rakotondravony & Radespiel, 2009;Rendigs et al., 2003;Thorén et al., 2011). M. murinus and M. ravelobensis live in partial sympatry (i.e., occurring together in some forests but alone in others) in the two study sites, exhibit similar body sizes (ca. 60 g), overlapping diets and seasonal reproductive activity Thorén et al., 2011;Zimmermann et al., 1998). According to previous studies (Kiene et al., 2020;Klein et al., 2018), seasonality, habitat fragmentation, host sex, body mass, and species seem to be important determinants of host ectoparasite infestations. However, the environmental factors driving parasite seasonality, as well as potential yearly variations in infestation patterns have not yet been investigated. We aim to test these and the potential effects of temperature and rainfall variations on the infestation dynamics of different ectoparasite taxa (temporary and permanent) on these two host species. Furthermore, other general factors that are expected to impact parasite dynamics (e.g., habitat, host species, sex, age, and body mass) will be considered. Over a period of 4 years (2010/2011, 2015/2016), all Microcebus that were captured and released between March and November each year were permanently marked, and many of them were repeatedly inspected for ectoparasites. Rainfall and temperature are explored as general predictors for the occurrence and richness of ectoparasites. Since high temperatures and low humidity can lead to desiccation and reduced ectoparasite survival (Berger et al., 2014;Kiene et al., 2020;Moyer et al., 2002;Sasa, 1961;Wall & Shearer, 1997), we predict an increase of ectoparasite infestations in years of lower (monthly) temperatures and in years of more rainfall (total or monthly). Moreover, due to their different ecology and life cycles, we predict that the impact of climatic variables will be higher for temporary parasites, such as ticks or chigger mites, that spend most of their life cycle in the environment, for example, as off-host larvae, nymphs, and adults in the vegetation or on the ground (Sasa, 1961;Wall & Shearer, 1997, 2001b, than for permanent parasites, such as sucking lice, which are more protected from outer influences by the stable microclimate on the body of their host.  (2010,2011,2015,2016) in the Ankarafantsika National Park in northwestern Madagascar (Supporting Information: Table S1). Data were collected from two separate study sites within the park. The first site, Jardin Botanique A (JBA; 16°19′07.2″ S, 46°48′35.5″ E), is a dry deciduous forest on sandy soil situated near the park headquarters at Ampijoroa at about 190 m elevation above sea level and with no surface water (Chanu et al., 2012). The second site, Jardin Botanique B (JBB, 16°18′02.6″ S, 46°48′44.7″ E), is a section of a gallery forest next to Lake Ravelobe at 89 m above sea level and is partly flooded during the rainy season (Rendigs et al., 2003;Sehen et al., 2010).

| METHODS
The climate in this region is strongly seasonal with a cool dry season from May to October and a hot rainy season from November to April (Figure 1). Ectoparasite sampling covered 3 months of the rainy season (March, April, and November) and all months of the dry season (May-October) (Supporting Information: Table S1). Between 126 and 219 individual mouse lemurs were examined per year, and both sexes were represented in both sites and both species (Table 1).
For the statistical analyses of host and temporal effects, consecutive ectoparasite examinations of a host individual in intervals of less than 7 days were merged using the following procedure: if there was one or more parasite detections in 7 days, then the lemur was recorded as positive for that ectoparasite, but if all were negative, then it was recorded as absent. In addition, inspection results with lacking body mass values were excluded. With this approach, the final number of sampling events used for modeling was 1940 ectoparasite inspection events stemming from 598 individuals. The term "infestation rate" is used whenever referring to a relative number of inspection events in which a specific ectoparasite type was found. By contrast, the term The presence or absence of the various ectoparasite types was recorded macroscopically in the field, and a small sample of all detected ectoparasite types was taken and stored in 90%-96% ethanol for each host individual and sampling day separately.
Collected ectoparasite samples were identified morphologically by applying standard identification methods under light microscopy to verify the field records (Durden et al., 2021;Kiene et al., 2020;Klein et al., 2018;Stekolnikov, 2018;Walker et al., 2003). Presence or absence was recorded for all ectoparasite taxa based on the composite record of field notes and microscopy for each mouse lemur examination. Positive field records for ectoparasites without a collected sample were considered only for ticks, because they cannot be easily overlooked or misidentified. If no field sample was available for other ectoparasite types (e.g., sucking lice, mites), positive records from the field were excluded (=not available) as their difficult macroscopic identification in the field could have led to false positive results.

| Data analysis
Statistical analyses were conducted for the three most M. murinus was not found in JBB across all years and months, the variable combining sampling site and host species was necessary to avoid modeling errors. Juvenile age was deduced from a low initial body mass (clearly below average) and its steady increase over the first months of capture . To control for pseudoreplication based on multiple sampling of single hosts, host identity ("animal ID") was added as a random factor. Since not all months were sampled in all years (Supporting Information: first subjected to a PCA that was based on a correlation matrix for all explanatory variables (after standardization of each entry as deviation from the mean, divided by SD). A total of nine uncorrelated principal components (PCs) resulted from the PCA, five of which, with an eigenvalue of >1, were selected for further analyses and explained 96.6% of the variance in the climatic data set (Supporting Information: Table S3). The factor coordinates of cases were used for the subsequent modeling steps and high factor loadings (>0.7 or <−0.7) of climatic variables were used for interpretation (Supporting Information: Table S4). The PCA was performed with STATISTICA 12 (Statsoft Inc.). The same four dependent variables as before (presence/absence of Haemaphysalis spp., S. microcebi, Lemurpediculus spp., EPSR) were analyzed by fitting GLMMs, with the first five PCs from the PCA used as fixed factors with the data set containing one entry per individual (N = 583). Preliminary models that contained more than one entry per individual and individual identity as an additional random factor did not converge and were therefore discarded. To distinguish between monthly variations due to seasonal ectoparasite dynamics and the effect of yearly climatic changes, month was added as a random factor.
All models were fitted with binomial assumption and logit-link for the presence-absence data, and Poisson assumption and with loglink for the EPSR. All calculations were done in Rstudio version 4.0.3 (Integrated Development for R. Rstudio Inc., http://www.rstudio. com). Models were fitted using the package "lme4" (Bates et al., 2015).
The selection of GLMMs for final interpretation was done for all global models using the automated model selection function dredge() of the R package "MuMln" (Barton, 2020), based on the Akaike Information Criterion (AIC) method (Burnham & Anderson, 2002).
The corrected AIC values (AICc) of models with all possible combinations of fixed factors were compared and all best models (ΔAIC < 2) were identified. When the factors year, month and site-species were significant in the best models, pairwise differences between levels were tested with a Tukey test (post hoc analysis), applying the R package "multcomp" (Hothorn et al., 2008). Graphics were obtained using the R package "ggplot2" (Wickham, 2016).

| Parasite identification and general infestation rates
The investigated hosts were infested by ticks (Acari, Ixodidae), sucking lice (Insecta, Anoplura), and mites of the families Trombiculidae, Laelapidae, and Atopomelidae (Supporting Information: Figure S1). Based on morphology, all ticks were identified as Haemaphysalis spp. According to Durden et al. (2018)     microcebi did not differ between sexes (Supporting Information: Table S6: both global models, p = n.s.). Significant effects of host age were neither found for the modeled ectoparasite taxa nor for EPSR (Table 4).
Body mass had a significant impact on all modeled ectoparasite taxa, but not on EPSR (

| Monthly and yearly dynamics in infestation risk
Significant effects of the variable month were detected in all best models and for all parasite taxa, but the effects differed between them (Table 4)  to October with an infestation peak in June (Figure 2b). Infestation rates were significantly higher in May, June, July, and August than in other months (Supporting Information: Table S6: global model A). Lemurpediculus spp. were recorded during the entire study period but with varying frequencies between months (Figure 2c).
GLMMs confirmed that infestation rates were significantly higher in October than in the other months except for March and November. Infestation rates in March, April, September, and November were also significantly higher than in May, June, July, and August (Supporting Information: Table S7: global model A).

The modeling of the EPSR revealed significantly higher values from
August to October in comparison to April, May, and June, but not compared to July and November (Supporting Information:  Figure 2d). Animals also showed significantly higher EPSR in July than in May.
Sampling year impacted the infestation rates of all three modeled parasite taxa and the EPSR significantly but in different ways (Table 4) Note: Table also includes an interpretation for each of the five principal components (in brackets), as derived from the climatic factors with high factor loadings (>0.7 or <−0.7, for details, see Letter coding (a-d) indicates significant differences (p < 0.05) between months in the best model (see Table 4 and Supporting Information: Tables S5-S8 for details). For example, months marked with letter "a" are significantly different from months carrying other letters. Letter coding (a-d) indicates significant differences (p < 0.05) between months in the best model (see Table 4 and Supporting Information: Tables S5-S8 for details). For example, months marked with letter "a" are significantly different from months carrying other letters.

| Effects of climatic variations on ectoparasite infestation risk
The first five PCs explained 96.61% of the total variance in the climate data set (Supporting Information: Table S3). PC1 and PC2 were best characterized by rainfall and temperature variations during the rainy season, while PC3 and PC5 showed the highest factor loadings for temperatures during the capture month and the month before capture, and PC4 represented a heterogeneous mix of all climatic factors (Supporting Information: Table S4). The PCA revealed strong climatic variations between the study years that were associated to rainfall and temperature variations during the previous rainy season (Supporting Information: Figure S2).
GLMMs revealed significant effects of PCs for all ectoparasite taxa and for EPSR, but the effects differed between them (Table 5, further details in Supporting Information: Tables S9-S12). Haemaphysalis spp. infestation rates were negatively associated with PC1 and positively with PC2 and PC3. Among these, PC3 (=warm temperatures during capture month and the month before capture) had the highest estimate and, therefore, the highest impact on the infestation risk with ticks (Supporting Information: Table S9). PC4 (=overall component) and PC5 (=warm nights during month before capture) were negatively related with S. microcebi infestations, among which warm night temperatures during the month before capture had the strongest negative effect (Supporting Information: Table S10). For Lemurpediculus spp., infestation rates were positively associated with PC1, but negatively associated with PC2, PC4, and PC5. Among these, the negative effect of PC5 (=warm night temperatures during the month before capture) showed the strongest impact (Supporting Information: Table S11). Finally, EPSR was significantly and negatively associated with PC1, PC3, PC4, and PC5, of which again PC5 had the strongest impact (Supporting Information: Table S12). Taken together, it appears that the temperatures during the capture month and the month before capture (PC3/PC5) consistently explain variations in infestation rates in all three ectoparasite taxa, although the relationship was positive in the case of Haemaphysalis spp., and negative for S. microcebi and Lemurpediculus spp.

| DISCUSSION
Although five ectoparasite taxa were identified on mouse lemurs in this study, only three of them (Haemaphysalis spp., S. microcebi, Lemurpediculus spp.) had sufficiently high occurrences to be used for the subsequent modeling approaches. While temporal, that is, monthly and yearly, variations of ectoparasite infestations were found in this study and will be discussed with the results on the impact of the climatic variables below, this study also revealed an impact of habitat type (=sampling site), and several host-related factors (species, sex, body mass) on the three most abundant ectoparasite taxa and on EPSR.

| Impact of study site and host-related factors (species, sex, age, and body mass)
Previous studies have shown that habitat quality, host species, sex, and body mass impact ectoparasite infestations in the two studied mouse lemur species (Kiene et al., 2020;Klein et al., 2018). However, the comparison of ectoparasite infestations in two study sites with largely different ecology but under the same climatic dynamics (Chanu et al., 2012;Sehen et al., 2010) allowed us to test the differential effect of species and habitat across multiple years, while controlling for sex, age, and body mass.
Our results showed that study site and host species affected the three parasite taxa differently. While Haemaphysalis (tick) infestation rates were neither affected by species nor study site, study site did affect S. microcebi (chigger), and both species and study site affected However, many tick species are not very host specific (Klompen et al., 1996), and Kiene et al. (2020) did not find differences in tick prevalences between the two mouse lemur hosts in a larger network of study sites in the same region. Likewise, the lack of study site differences for ticks is unexpected, since JBB is more humid than JBA and should, therefore, provide a more favorable environment for offhost tick survival than the drier JBA. Further studies on the microhabitat requirements of the environmental stages of this parasite, that is, by sampling tick stages directly in the forest environment, and more detailed characterization of the specific environmental conditions in each study site will be needed to better understand its biology.
S. microcebi infestation rates were higher in M. ravelobensis inhabiting JBB than those from JBA, and higher than in M. murinus from JBA. Neither in JBA, nor in JBB, were differences between host species found. These results support an effect of study site but not of host species for this ectoparasite taxon. Since trombiculid mites are temporary parasites that spend most of their life cycle off the host and in leaf litter or soil (Sasa, 1961;Shatrov & Kudryashova, 2006), it is not surprising that study site, that is, the different habitat conditions in JBA and JBB, significantly affected its occurrence. In tropical regions, it has been shown that trombiculid mites are limited by precipitation and humidity (Sasa, 1961). Although precipitation does not differ between JBA and JBB as they are only 3 km apart, the forest in JBB is generally more humid than JBA due to its proximity to Lake Ravelobe (Sehen et al., 2010). This aspect could make JBB a more suitable habitat for S. microcebi than JBA. Future comparative studies on the microclimate in each study site and the microhabitat requirements of the environmental stages of this chigger will be needed to fully understand the causal abiotic and biotic factors that drive spatial variations in mouse lemur infestations with trombiculid mites.
Concordant with previous findings for JBA during the 2015/ 2016 field season, Lemurpediculus spp. infestation rate was higher in M. ravelobensis than in M. murinus, which was previously explained with a higher degree of group sleeping and different composition of sleeping groups of M. ravelobensis compared to M. murinus, at least during those study years (Klein et al., 2018). In contrast to S. microcebi, infestation with Lemurpediculus lice was higher in JBA than in JBB, which may have resulted from differences in host population density between JBA and JBB. Many studies relate high host population densities to higher transmission rates, higher parasite prevalence, and higher parasite species richness (Arneberg, 2002;Krawczyk et al., 2020;Stringer & Linklater, 2015). A recent study from JBB suggested a substantial population decline for M.
ravelobensis in JBB between 2010 and 2016 based long-term trapping data (Henkel et al., 2019(Henkel et al., /2020, which may have led to low louse transmission rates, and as a consequence, to low infestation rates in JBB. Host sex has been shown to influence ectoparasite infestation susceptibility in many taxa (Fagir et al., 2015;de Mendonça et al., 2020;Zohdy et al., 2017) including Similarly, Kiene et al. (2020) found no effect of host sex on tick and mite infestations, but confirmed such an effect for sucking lice infestation, being higher in males than in females. Our study, which integrates and expands the datasets of both Durden et al. (2021) and Klein et al. (2018), did indeed detect higher infestation rates in male than in female mouse lemurs for Haemaphysalis spp., Lemurpediculus spp. and for EPSR. Whether these differences are due to testosterone-mediated differences in immunity or to sex-specific patterns of behavior, seasonal ranging and, therefore, exposure to these parasites (Klein, 2004;Muehlenbein & Watts, 2010;Waterman et al., 2014;Zohdy et al., 2012) cannot be clarified without further detailed integrative studies.
In our study, body mass was negatively associated with Lemurpediculus spp. infestation, but positively associated with Haemaphysalis spp. and S. microcebi prevalences. Given that the body mass of mouse lemurs was previously shown to decrease significantly over the course of the dry season (Klein et al., 2018), when food availability is low and reproductive activities start , it is not possible to disentangle the drivers of the negative relationship between body mass and sucking lice infestation, in particular, since they are nonexclusive.
The positive relationship between body mass and Haemaphysalis spp. and S. microcebi infestation shown in this study were not reported in the previous studies by Klein et al. (2018) and Kiene et al. (2020). Given that host body mass significantly impacted infestation rates for ticks and mites only when the study year was not part of the model, the influence of body mass may indirectly reflect year-to-year variations in body condition than, for example, choosiness of temporary parasites for a well-fed host (Christe et al., 2003).

| Monthly dynamics
All three analyzed parasite taxa in this study showed persistent monthly variations in infestation patterns. Haemaphysalis ticks were found most frequently from April to November and had an infestation peak in August. The great majority of the collected specimen were larvae or nymphs, with only 15 adult Haemaphysalis spp. (of a total of more than 700 ticks) being found on Microcebus (Klein et al., 2018;personal observation). These results are congruent with previous findings by Klein et al. (2018) showing that tick infestation prevalences on mouse lemurs increased over the dry season, when food resources are less abundant (Thorén et al., 2011), and lemurs may need to descend more often to the floor while foraging. This might increase exposure to these parasites, which are usually in the lower vegetation questing for hosts (Walker et al., 2003).
Low infestation intensities during the months March, April, and May might be related to the univoltine life cycle of Haemaphysalis spp. and its highly seasonal development, as described by Rodriguez et al. (2015). Klein et al. (2018) also suggested that adult ticks may mostly occur during the rainy season, when high humidity provides a favorable environment for egg development and hatching success (Randolph, 2004). Further, infestation rates were (at least partly) associated with warmer temperatures during the capture month and the month before capture, which might reflect that ticks need to find a host to avoid desiccation (Benoit & Denlinger, 2010;Randolph, 2004;Walker et al., 2003).
Since the only parasitic stage in chigger mites is the larva, these mites spend most of their life cycle in the environment (Mullen & O'Connor, 2019;Sasa, 1961;Shatrov & Kudryashova, 2006;Wall & Shearer, 2001a). In our study, S. microcebi chiggers were found on mouse lemurs mainly from April to October, with an infestation peak in June. Since egg development is likely the most susceptible stage to desiccation, eggs are probably deposited during the rainy season, and hence larvae should increasingly search for hosts over the first months of the dry season (Sasa, 1961 It has been stated that the development and survival of chiggers in tropical environments depends mostly on humidity and rainfall (Sasa, 1961;Wall & Shearer, 2001a). These mites spend most of their life cycle in the leaf litter or soil, and their successful reproduction, parts of their development and survival depend on specific environmental conditions in that particular forest stratum (Schöler, 2003). In our study, S. microcebi infestation rates were highest in 2011 and lowest in 2010 and 2015. Our expectation of higher infestations under lower temperatures and in years of more rainfall was only partially supported. Only one climatic variable was clearly identified impacting chigger infestation risk significantly: infestation risk increased with lower overnight temperatures during the month before capture (=PC5). Because corresponding data on humidity, soil temperatures, and moisture in the study sites are not currently available, and we lack knowledge of the environmental requirements of the different stages of S. microcebi, causal explanations for abiotic effects on S. microcebi survival, reproduction, and resulting infestation risks for hosts cannot be currently identified.
Moreover, complex indirect effects, such as impacts of climatic variations on plant phenology and leaf litter, may also shape conditions in the microhabitats of these parasites (Anu et al., 2009) and require further investigation.
We expected the directly transmitted (host to host) sucking lice to be least impacted by year-to-year environmental changes, although it has previously been shown that UV radiation and aridity might have negative consequences for sucking lice development and survival (Durden & Musser, 1994;Kiene et al., 2020;Moyer et al., 2002;Wall & Shearer, 2001a). Our modeling results revealed that Lemurpediculus spp. showed distinct yearly variations, with highest infestation rates in 2016, followed by 2010/2015, and then 2011. Therefore, and in contrast to our expectations, sucking louse infestation risk was highly impacted by climatic variations, both by variations in rainfall and temperature during the rainy season (PC1, PC2), and by temperatures during the months during and before host capture (PC5). Whether sucking lice responded to these environmental changes directly or indirectly, that is, if the environmental changes impacted the lice directly by modifying their microenvironment in the fur of the hosts or by impacting the ecology (e.g., feeding strategies, shelter use) and behavior of the host species and in consequence modulated social encounter, parasite removal or transmission opportunities for sucking lice, needs to be investigated in future studies.
Taken together, this study shows that several host-related their seasonal ecosystems. This is more relevant than ever in times of ongoing and strong anthropogenic pressures and alarming climate change scenarios that are already affecting Madagascar with its highly fragmented and vulnerable forest remnants (Ingram & Dawson, 2005;Vieilledent et al., 2018).